fbpx
Connect with us

The Conversation

AI supercharges data center energy use – straining the grid and slowing sustainability efforts

Published

on

theconversation.com – Ayse Coskun, Professor of Electrical and Computer Engineering, Boston University – 2024-07-11 07:26:28
A data center in Ashburn, Va., the heart of so-called Data Center Alley.
AP Photo/Ted Shaffrey

Ayse Coskun, Boston University

The artificial intelligence boom has had such a profound effect on big tech companies that their energy consumption, and with it their carbon emissions, have surged.

The spectacular success of large language models such as ChatGPT has helped fuel this growth in energy demand. At 2.9 watt-hours per ChatGPT request, AI queries require about 10 times the electricity of traditional Google queries, according to the Electric Power Research Institute, a nonprofit research firm. Emerging AI capabilities such as audio and video generation are likely to add to this energy demand.

The energy needs of AI are shifting the calculus of energy companies. They’re now exploring previously untenable options, such as restarting a nuclear reactor at the Three Mile Island power plant that has been dormant since the infamous disaster in 1979.

Advertisement

Data centers have had continuous growth for decades, but the magnitude of growth in the still-young era of large language models has been exceptional. AI requires a lot more computational and data storage resources than the pre-AI rate of data center growth could .

AI and the grid

Thanks to AI, the electrical grid – in many places already near its capacity or prone to stability challenges – is experiencing more pressure than before. There is also a substantial lag between computing growth and grid growth. Data centers take one to two years to build, while adding new power to the grid requires over four years.

As a recent from the Electric Power Research Institute lays out, just 15 states contain 80% of the data centers in the U.S.. Some states – such as Virginia, home to Data Center Alley – astonishingly have over 25% of their electricity consumed by data centers. There are similar trends of clustered data center growth in other parts of the world. For example, Ireland has become a data center nation.

AI is a big impact on the electrical grid and, potentially, the climate.

Along with the need to add more power generation to sustain this growth, nearly all countries have decarbonization goals. This means they are striving to integrate more renewable energy sources into the grid. Renewables such as wind and solar are intermittent: The wind doesn’t always blow and the sun doesn’t always shine. The dearth of cheap, green and scalable energy storage means the grid faces an even bigger problem matching supply with demand.

Advertisement

Additional challenges to data center growth include increasing use of water cooling for efficiency, which strains limited fresh water sources. As a result, some communities are pushing back against new data center investments.

Better tech

There are several ways the industry is addressing this energy crisis. First, computing hardware has gotten substantially more energy efficient over the years in terms of the operations executed per watt consumed. Data centers’ power use efficiency, a metric that shows the ratio of power consumed for computing versus for cooling and other , has been reduced to 1.5 on average, and even to an impressive 1.2 in advanced facilities. New data centers have more efficient cooling by using water cooling and external cool when it’s available.

Unfortunately, efficiency alone is not going to solve the sustainability problem. In fact, Jevons paradox points to how efficiency may result in an increase of energy consumption in the longer run. In addition, hardware efficiency gains have slowed down substantially, as the industry has hit the limits of chip technology scaling.

To continue improving efficiency, researchers are designing specialized hardware such as accelerators, new integration technologies such as 3D chips, and new chip cooling techniques.

Advertisement

Similarly, researchers are increasingly studying and developing data center cooling technologies. The Electric Power Research Institute report endorses new cooling methods, such as air-assisted liquid cooling and immersion cooling. While liquid cooling has already made its way into data centers, only a few new data centers have implemented the still-in- immersion cooling.

a man wearing rubber gloves and a visor lowers a circuit board into a trough containing a liquid
Running computer servers in a liquid – rather than in air – could be a more efficient way to cool them.
Craig Fritz, Sandia National Laboratories

Flexible future

A new way of building AI data centers is flexible computing, where the key idea is to compute more when electricity is cheaper, more available and greener, and less when it’s more expensive, scarce and polluting.

Data center operators can convert their facilities to be a flexible load on the grid. Academia and industry have provided early examples of data center demand response, where data centers regulate their power depending on power grid needs. For example, they can schedule certain computing tasks for off-peak hours.

Implementing broader and larger scale flexibility in power consumption requires innovation in hardware, software and grid-data center coordination. Especially for AI, there is much room to develop new strategies to tune data centers’ computational loads and therefore energy consumption. For example, data centers can scale back accuracy to reduce workloads when training AI models.

Realizing this vision requires better modeling and forecasting. Data centers can try to better understand and predict their loads and conditions. It’s also important to predict the grid load and growth.

Advertisement

The Electric Power Research Institute’s load forecasting initiative involves activities to with grid planning and operations. Comprehensive monitoring and intelligent analytics – possibly relying on AI – for both data centers and the grid are essential for accurate forecasting.

On the edge

The U.S. is at a critical juncture with the explosive growth of AI. It is immensely difficult to integrate hundreds of megawatts of electricity demand into already strained grids. It might be time to rethink how the industry builds data centers.

One possibility is to sustainably build more edge data centers – smaller, widely distributed facilities – to bring computing to local communities. Edge data centers can also reliably add computing power to dense, urban regions without further stressing the grid. While these smaller centers currently make up 10% of data centers in the U.S., analysts the market for smaller-scale edge data centers to grow by over 20% in the next five years.

Along with converting data centers into flexible and controllable loads, innovating in the edge data center may make AI’s energy demands much more sustainable.The Conversation

Ayse Coskun, Professor of Electrical and Computer Engineering, Boston University

Advertisement

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Read More

The post AI supercharges data center energy use – straining the grid and slowing sustainability efforts appeared first on .com

Advertisement

The Conversation

‘Difficult’ children are only slightly more likely to have insecure attachments with parents

Published

on

theconversation.com – Or Dagan, Professor of Clinical Psychology, Long Island Post – 2024-09-11 07:31:47

may worry about connecting with a child who is hard to comfort.
hobo_018/E+ via Getty Images

Or Dagan, Long Island University Post and Carlo Schuengel, Vrije Universiteit Amsterdam

with difficult temperaments, personality tendencies such as irritability and a hard time being comforted, are only slightly more likely than other children to have insecure attachment relationships with one or both of their parents, according to our research. This finding refutes the long-standing notion held by many psychologists that early attachment behaviors are mainly determined by a child’s temperament.

An attachment relationship reflects the child’s expectations about their caregiver in times of need. A secure attachment is likely if a caregiver is consistently available and emotionally supportive when the child is alarmed. However, if a child learns that their caregiver will not be there when needed or will not effectively soothe them, chances are they will develop an insecure attachment relationship with that caregiver.

Advertisement

As researchers in clinical psychology and child and family studies, we are interested in how the quality of child-caregiver relationships affects children’s .

Part of that is understanding what influences the way child-parent attachments form. Even infants show stark differences in temperament, and some psychologists have argued that these individual dispositions may also explain how youngsters interact with caregivers. Even the most sensitive and loving parent could feel challenged when caring for a difficult child – does that dynamic influence the quality of their attachment relationship?

How we do our work

With 29 other researchers, we started a research consortium to study the quality of children’s attachment relationships with their mothers and fathers – or what we call their attachment networks.

For this meta-analysis, we combined data collected over the past 40 years on 872 children from North American families.

Advertisement

Researchers observed these kids interacting separately with their mothers and fathers and completed assessments designed to evaluate children’s attachment behaviors: How do they seek comfort in times of need? How easily do they find reassurance in their parents? How do they explore their in the presence of their parents?

In addition, parents reported on the degree of their children’s difficult temperament. How likely was the child to experience intense negative emotions, such as anger, sadness or fearfulness?

It seemed logical that children who have a more difficult temperament might tend to have more insecure attachment relationships within the – but that’s not exactly what we found. Instead, a difficult temperament had very little to do with the number of insecure attachment relationships a child had with their parents.

Greater tendencies toward a difficult temperament had just a very small impact – less than 1%, according to our statistical analysis – on children’s likelihood to have multiple insecure attachment relationships. And temperament was only slightly more difficult in infants who had insecure attachment with both parents rather than with only one or neither of the parents.

Advertisement
man kneels comforting girl holding a cone with her ice cream on the ground
Difficult temperament doesn’t mean a secure attachment relationship won’t form between child and parent.
PeopleImages/iStock via Getty Images Plus

Why it matters

Recent research from our consortium has found that children who develop secure attachment relationships with both mom and dad tend to show fewer symptoms of anxiety and depression, as well as stronger language skills, compared with those who had only one or no secure attachment relationships in their two-parent families.

The results from our latest study suggest that even children with inborn characteristics of a difficult temperament can benefit from the advantages that from multiple secure attachments. These findings may reassure worried parents.

What still isn’t known

Further research is needed to explain the small increase in insecure attachments we did identify for kids with difficult temperaments.

For example, children’s difficult temperament increases the likelihood of negative parenting, including anger and coercing and overcontrolling behaviors. These in turn can intensify children’s negative emotions. Over time, it’s possible this two-way street of negative reactions increases the probability of insecure attachment relationships between child and parents.

Interventions that promote positive parenting and sensitive discipline increase children’s attachment security with their parents. So, parents can foster secure attachment with deliberate efforts to be more sensitive to the child’s emotional needs.

Advertisement

One thing has become clear. The secure attachment relationships that play a critical role in cognitive and emotional development are not closed off to children born with difficult temperaments.

The Research Brief is a short take on interesting academic work.The Conversation

Or Dagan, Professor of Clinical Psychology, Long Island University Post and Carlo Schuengel, Professor of Clinical Child and Family Studies, Vrije Universiteit Amsterdam

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Read More

Advertisement

The post ‘Difficult’ children are only slightly more likely to have insecure attachments with parents appeared first on .com

Continue Reading

The Conversation

With China seeking AI dominance, Taiwan’s efforts to slow neighbor’s access to advanced chips needs support from the West

Published

on

theconversation.com – Min-Yen Chiang, PhD student in political science , Georgia University – 2024-09-11 07:33:43

Taiwanese Vice President Lai Ching-te gives a speech at the CommonWealth Semiconductor Forum in 2023 in Taipei, Taiwan.
Annabelle Chih/Getty Images

Min-Yen Chiang, Georgia State University and Robert Muggah, Pontifícia Universidade Católica do Rio de Janeiro (PUC-Rio)

Tensions between China, Taiwan and the U.S. aren’t limited to aerial military maneuvers and drills on the high seas. The shadow conflict is also playing out in the technological arena.

One of the central drivers of the deepening geopolitical rifts between China on one side and Taiwan and the U.S. on the other is dominance over global semiconductor supply chains. This is because semiconductors – or microchips – power everything from smartphones and home office software to critical and advanced military hardware.

Advertisement

As international demand for sophisticated microchips surges, not least owing to the blistering growth of artificial intelligence, so does their strategic value to the global and the progress of individual nations. China today spends as much importing microchips as it does importing oil.

This deepening reliance on semiconductors around the world adds another layer of complexity to simmering China-Taiwan tensions. Today, Taiwan is the world’s largest and most advanced microchip producer, and China is the planet’s biggest consumer of semiconductors.

As researchers in geopolitics and advanced technologies, we see the competition to control microchip supply chains as one of the defining struggles of the 21st century. Taiwan’s experience could serve as an example to the U.S., which on Sept. 6, 2024, announced a fresh wave of export controls on semiconductor goods.

The world’s chipmaker

Taiwan did not emerge as the world’s semiconductor powerhouse by accident. The self-governing island has been producing high-quality microchips for decades due in large part to its flexible production network and world-class engineering talent pool.

Advertisement

Yet Taiwan faces a delicate balancing act in maintaining its market superiority in semiconductors, especially when it to exporting advanced technologies to China. For one, Taiwanese policymakers are understandably determined to both avoid political entanglements with a country that views the island as its own territory and hold on to the island’s intellectual property. Moreover, Taiwan wants to keep microchips from powering Chinese missiles currently pointed at the capital, Taipei.

The road to regulating chips

Until the early 1990s, the transfer of technologies to China was prohibited under Taiwanese law. But regulations were weakly enforced. As a result, Taiwanese businesses frequently circumvented existing sanctions by rerouting investments through then-British Hong Kong. The reality was that the chip industry was a lucrative source of revenue for the island.

Taiwan’s approach to regulating the flow of technologies started to change in 1993 when President Lee Teng-hui implemented the “no haste, be patient” policy. The strict ban was relaxed and replaced by a system in which additional layers of oversight were added to highly advanced technologies, deals valued at more than US$50 million and specialized critical infrastructure projects.

Advertisement

Crafted over decades, this “outbound investment screening” system features multiple checks intended to safeguard Taiwan’s core chip technologies. Taiwanese authorities are actively involved in monitoring and overseeing investment decisions involving China made by the island’s semiconductor companies. are also keen to ensure that local chipmakers are aligned with Taiwan’s strategic interests, while minimizing political ties with its neighbor.

During the screening , Taiwanese companies are required to submit detailed investment plans to -appointed reviewers for approval. For example, when a Taiwanese semiconductor firm, such as the world’s largest chip manufacturer TSMC, considers establishing a new facility in China, it must first undertake a rigorous approval process.

Changing calculations

While the cautious policy shift appears prescient today given rising geopolitical tensions, at the time it was considered out of step with the direction of more open global trade relations with China. The restrictive human rights considerations that had curbed Western trade with China were eased in the 1990s after intensive lobbying by U.S. corporations. In 2000, U.S. President Bill Clinton granted China permanent normal trade relations, paving the way for its accession to the World Trade Organization a year later. Trade with China, of advanced technologies, exploded thereafter.

The silhouette of a person is seen in front of a sign reading 'TSMC'
A visitor explores the TSMC exhibition at the World Semiconductor Congress 2022 in Nanjing, Jiangsu province, China.
CFOTO/Future Publishing via Getty Images

But Washington’s strategic calculations over trade with China have shifted dramatically over the past decade. In 2018, the U.S. singled out China as a strategic competitor, designating several Chinese hackers and the government itself as national security threats. By August 2023, President Joe Biden directed the Treasury Department to draft regulations to develop an outbound investment security program to safeguard semiconductor, quantum and AI technologies.

A few months later, the U.S. issued sweeping restrictions on the trade of advanced chips and chipmaking equipment with China. In early 2024, the European Union released a white paper proposing to do the same.

Advertisement

Of course, Taiwan has its own specific political concerns when it comes to China. Given Beijing’s long-standing ambition to, as Chinese put it, “reunify” Taiwan with the mainland, local officials are particularly aware how doing business with China might have unpredictable and damaging political ramifications.

The Taiwanese National Security Bureau has long warned that Beijing is using business to covertly advance its political ambition, including by leveraging Taiwanese capital to build influence and proxies within Taiwan. And in late 2023, Taiwan’s National Science and Technology Council announced a list of over 20 core technologies it wanted to prevent Beijing from acquiring, including know-how and raw material to make chips smaller than 14 nanometers.

New challenges for Taiwan’s regulations

Taiwanese authorities and businesses have built on the outbound screening system in order to push back against Chinese influence. In recent years, additional principles to protect Taiwan’s semiconductor dominance have been introduced, including requiring Taiwanese investors to retain a controlling interest in all Chinese subsidiaries.

Nonetheless, Taiwan’s outbound investment screening system is facing multiple tests. While it is designed to curb the transfer of advanced Taiwanese technologies to China, it also has to oversee financial investments from Taiwan into China’s surging chipmaking sector.

Advertisement

In 2022, for example, the Taiwanese technology group Foxconn announced an investment in Tsinghua Unigroup through its Chinese subsidiary. Tsinghua Unigroup is backed by China’s National Integrated Circuit Industry Investment Fund and controlled by a Beijing-based private equity firm. Owing to Foxconn’s failure to submit a required preapproval application to the outbound investment screening authorities, the Taiwanese government imposed a fine on the company, which eventually withdrew its investment.

A person in a hazmat suit walks through a lab.
Inside the Taiwan Semiconductor Research Institute in Hsinchu, Taiwan.
Annabelle Chih/Getty Images

China’s growing chip industry is also expanding its local supply chain, raising questions about whether Taiwan should expand restrictions on other suppliers linked to semiconductor manufacturers. After the U.S. introduced export controls on China in late 2023, the Chinese firm Huawei aggressively expanded its chip production network by leveraging its affiliates and Taiwanese suppliers. Four Taiwanese semiconductor firms that had previously been approved for outbound investment were subsequently accused of aiding Huawei in building China’s domestic chip supply chain.

Confronting China’s ambition

With access to Taiwanese semiconductors increasingly restricted, China has aggressively pursued greater technological autonomy. It has done so by reducing its reliance on imports of advanced equipment and materials from U.S., Japan, the Netherlands and Taiwan.

There are legitimate concerns in the West that tightening international export restrictions on microchips and relevant suppliers could inadvertently strengthen China’s determination to accelerate the of its domestic semiconductor production.

Official data appears to corroborate this view; China’s overall imports of microchips in 2023 were below 2017 levels. Exports of Taiwanese chips to China dropped by 18% in 2023.

Advertisement

Meanwhile, China’s National Bureau of Statistics reported that overall domestic chip production grew by 40% in the first quarter of 2024. Its share of global capacity to produce logic chips at 10-22 nanometers could rise from 6% to 19% by 2032.

But these data points do not necessarily mean that China is close to technological autonomy. Most of the increases in domestic chip production involve “mature” chips for household appliances and electric vehicles, rather than the most advanced chips required to accelerate AI computing power.

Meanwhile, China is still dependent on Taiwan for its semiconductors. The decrease in overall chip imports could be a result of international export restrictions on the most cutting-edge semiconductors needed for high-end smartphones and other AI-driven, high-performance computing products.

Coordinating international efforts

Restricting China’s access to the global superconductor supply chain is challenging. While doing so makes China reliant on Taiwanese chips – and as such may serve as a temporary protective shield against invasion – it could also exacerbate Beijing’s insecurities, pushing President Xi Jinping to hasten efforts to become technologically self-sufficient in advanced chips manufacturing. At the same time, outright bans on these chips hasn’t prevented China from producing a range of semiconductors using foreign capital and technology.

Advertisement

To address this challenge, Taiwan’s screening mechanisms not only need to remain nimble and vigilant – they need to be supported by a coordinated international approach. Only then will it be possible to slow the progress of authoritarian regimes in the AI race.The Conversation

Min-Yen Chiang, PhD student in political science , Georgia State University and Robert Muggah, Lecturer, Pontifícia Universidade Católica do Rio de Janeiro (PUC-Rio)

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Read More

The post With China seeking AI dominance, Taiwan’s efforts to slow neighbor’s access to advanced chips needs support from the West appeared first on theconversation.com

Advertisement
Continue Reading

The Conversation

How we discovered that people who are colorblind are less likely to be picky eaters

Published

on

theconversation.com – Isabel Gauthier, David K. Wilson Professor of Psychology, Vanderbilt – 2024-09-10 07:28:58

How we discovered that people who are colorblind are less likely to be picky eaters

Watching Julia Child in color or black and white could influence how appetizing the food looks to some audience members.
AP Photo

Isabel Gauthier, Vanderbilt University

The seventh season of Julia Child’s “The French Chef,” the first of the television to in color, revealed how color can change the experience of food. While Child had charmed audiences in black and white, seeing “Bouillabaisse à la Marseillaise” in color helped elevate the experience from merely entertaining to mouthwatering.

I am a psychologist who studies visual abilities. My work, through a serendipitous research journey into individual differences in food recognition, uncovered a unique role for color in emotional responses to food.

Advertisement
three versions of an image of a bowl of seafood in broth
Bouillabaisse à la Marseillaise, in gray scale, color, and simulating the most common form of color blindness (based on daltonlens.org).
Image Professionals GmbH/Foodcollection via Getty Images

People vary in their ability to recognize food

This journey started when my and I measured how people vary in their ability to recognize images of prepared food. Over the past 20 years, we and other researchers have learned that people vary more than originally suspected in how well they discriminate and identify objects, like birds, cars or even faces.

It seems obvious that some people know more than others about birds or cars. Yet, interestingly, there is as much variation in face recognition ability, even though virtually every sighted person has experience seeing faces.

Experience with food is also universal. We were curious how much people would vary in their ability to recognize food items. Our tests simply ask people to match images of the same dish among similar ones, or to find the oddball dish among others. People vary a great deal on these tasks, and some of this variation is explained by a general ability to recognize objects of any kind.

But a portion of the food recognition differences among people was not explained by this general ability. Instead, we speculated that this variability may be related to people’s attitude toward new foods. People who strongly endorse statements like “I don’t trust new foods” or “I am very particular about the foods I eat” have what’s called food neophobia. It can lead to poor diet quality, resulting in nutritional deficiencies and higher risk of chronic disease.

As we predicted, we found that picky eaters scored worst on our tests of food recognition: Food neophobia is negatively correlated with food recognition ability.

Advertisement
pictures of four menu items side by side, three show various presentations of a caprese salad with mozzarella and basil, one shows a pizza with mozzarella and basil
Example of a test trial to measure food recognition ability: Which of these foods is the oddball? Top row shows the dishes in color, while the bottom row simulates the most common form of color blindness (based on daltonlens.org). The second image is the oddball: Notice how the pizza crust is more obvious with normal color.
Rouzes/E+, Carlo A/Moment, Yulia Naumenko/Moment, rudisill/E+ via Getty Images

Color connects food neophobia to recognition

While we were publishing our results, other scientists were debating new findings about how the brain reacts to food and color. Different research groups had identified brain in the visual system that responded preferentially to images of food. For instance, looking at a bowl of pasta would activate these brain areas, but not looking at a pile of string.

The scientific disagreement was about what it meant to identify a selectivity for food in areas of the brain already known for their responsiveness to color.

One group proposed that these parts of the brain responded to color because they are specialized to recognize food. The other group argued that color was not critical to the brain’s response to food. They even showed they could get similar brain activation when people looked at gray-scale images of food.

Was it possible that color was not critical to food recognition, while still playing a special role? We decided to replicate our initial study, with images of food in gray scale.

The results could have been predicted by Child herself: Without color, people unsurprisingly made a few more errors and mistook different dishes as the same kind of food, but the pattern of variation across people was otherwise unchanged. Those with a higher general visual ability did better with food, and we again found a specific ability for food that goes beyond this general effect.

Advertisement

But we did find one effect of removing color: food neophobia was no longer correlated with food recognition ability. It was as if whatever advantage the adventurous eaters had gained over picky eaters was all dependent on color.

Based on these results, we proposed two separate components of food-specific recognition ability. One is independent of color and explains why the results are the same in the experiments with and without color. The other one, related to emotional responses, is based on color and evidenced by the finding that food neophobia is only related to food recognition when the food appears in color.

We then made an entirely new prediction: Would people with color blindness – , really, because color blindness affects 16 times more men than women – be less food neophobic than those with normal color perception? Because someone who is colorblind experiences food in a restricted range of color, some of the signals that raise flags about freshness, safety or otherwise drive anxiety about novel food could be limited.

We recruited participants online, men in our study based on how they answered one question about color blindness that was buried in a long screening questionnaire. Our participants had no idea that we were interested in color blindness when we asked them then to fill out the Food Neophobia Scale that measures how resistant people are to new foods.

Advertisement

We found that colorblind men were indeed less food neophobic than non-colorblind men. We replicated this finding in another study, with colorblind men also reporting lower levels of food disgust. Seeing the world with a restricted color palate seems to mitigate an emotion-based resistance to new foods.

Blackberries in various stages of ripeness on the bush, from pale pink to red to dark black
The color of foods can encode important information.
Steve Goossen/Design Pics via Getty Images

Color and emotional responses to food

Our research is broadly consistent with other findings. Color can you decide whether food is cooked or if produce is ripe or rotten, and people tend to prefer meals with a range of colorful foods. Other research shows that color can influence what food tastes like. Some biologists have argued that the coloration of plants and the ability of animals to detect it have co-evolved.

The role of color in emotional responses to food up new avenues for addressing extreme cases of food neophobia. As researchers learn more about the intricate relationship between color perception and food, we might develop targeted interventions to improve dietary habits. Just like Julia Child’s colorful dishes, understanding and leveraging the power of color could enhance the appreciation and enjoyment of food.The Conversation

Isabel Gauthier, David K. Wilson Professor of Psychology, Vanderbilt University

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Read More

Advertisement

The post How we discovered that people who are colorblind are less likely to be picky eaters appeared first on .com

Continue Reading

Trending